New synthetic methodology for ring expansion of n-sized conjugated cycloalkenones into homoallylic n+3 lactones: Three-step synthesis of fragrant phoracantholide

Article abstract of DOI:10.1016/S0040-4039(00)01739-1

Under the influence of Lewis acidic boron trifluoride-diethyl etherate, epoxides are activated sufficiently to undergo rapid nucleophilic opening by ketone enolate ions at -78°C, forming useful γ-lactols. Application of this protocol to n+3 ring expansion features one-flask, multicomponent, sequential reactions. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01739-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 9655–9659
New synthetic methodology for ring expansion of n-sized
conjugated cycloalkenones into homoallylic n+3 lactones:
three-step synthesis of fragrant phoracantholide†
Gary H. Posner,* Qiang Wang, Bethany A. Halford, Jeffrey S. Elias and John P. Maxwell
Department of Chemistry, School of Arts and Sciences, The Johns Hopkins University, Baltimore, MD 21218, USA
Received 31 August 2000; revised 2 October 2000; accepted 3 October 2000
Abstract
Under the influence of Lewis acidic boron trifluoride–diethyl etherate, epoxides are activated sufficiently
to undergo rapid nucleophilic opening by ketone enolate ions at −78°C, forming useful g-lactols.
Application of this protocol to n+3 ring expansion features one-flask, multicomponent, sequential
reactions. © 2000 Elsevier Science Ltd. All rights reserved.
Ketone enolate anions couple with a large variety of electrophiles, forming new car-
bonꢀcarbon bonds in products of increased structural complexity.1,2 In contrast to ester and
amide enolate ions, however, ketone enolate ions do not generally open epoxides.3 Crotti and
colleagues have shown that, among diverse Lewis acids, scandium triflate activates epoxides
most effectively toward intermolecular nucleophilic attack by ketone enolates, albeit requiring
long reaction times (]18 h) at room temperature.4,5 We report here our recent discovery that
boron trifluoride–diethyl etherate activates some 1-substituted and 1,2-disubstituted epoxides
toward nucleophilic opening by ketone enolates under very mild reaction conditions. Examples
of such syntheses of g-hydroxy ketone products (Eq. (1)) and intermediates (Eq. (2)) are
presented here, illustrating high diastereoselectivity (Eq. (1)) and a new, two-step, stereocon-
trolled, n+3 ring expansion methodology (Eq. (2)).6
H
HO
O
O
H
1) LiN(SiMe3)2, THF, DME
H
(1)
2)
O
, BF3-OEt2
1, 80%, 89:11 dr
-78 °C, 0.5 h
* Corresponding author.
†
Dedicated to Harry Wasserman with respect, admiration, and friendship on the happy occasion of his 80th
birthday.
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01739-1

9656
H
H
HO
O
O
6
O
HO
1) LiSnMe3
O
b
b
H
H
2)
a
, BF3-OEt2
SnMe3
a
SnMe3
-78 °C, 5 min
2a
2b
Pb(OAc)4
CaCO3
benzene, reflux
(2)
OAc
OAc
AcO
O
Pb
H
O
O
H
O
–Pb(OAc)2
Me3Sn
AcO
H
9
H
3, 50% overall yield
The highly diastereoselective and mild formation of g-hydroxy ketone 1 using BF3·OEt2 as
Lewis acid (Eq. (1)) contrasts sharply with the low diastereoselectivity reported for similar ketone
enolate opening of epoxides catalyzed by Sc(OTf)3.4
We have chosen initially to establish the generality of Eq. (2) that showcases a one-flask, 3-
component, sequential process in which conjugate addition of trimethyltinlithium to 2-cyclo-
hexenone generated an enolate ion that nucleophilically opened boron trifluoride-complexed
cyclohexene oxide within 5 min at −78°C; the crude mixture of g-hydroxy ketone 2a and the cor-
responding stannyl hemiketal 2b was then fragmented by lead tetraacetate7 into 9-membered
olefinic lactone 3 in 50% overall yield. Formation of only the E-olefinic double bond in homoal-
lylic product lactone 3 and the established concerted nature of the lead tetraacetate fragmentation
mechanism8 indicate that, as expected, the two new bonds a and b in hemiketal 2b are trans to
each other. In Table 1 are summarized results for this two-step, multicomponent ring expansion
of 6- and 7-membered conjugated cycloalkenones into 3-atom enlarged, regiospecifically and
stereospecifically unsaturated, 9- and 10-membered lactones. These transformations represent the
synthetic equivalent of n+3 homologous Baeyer–Villiger oxidations9 in which the original
cycloalkenone Z-double bond is inverted into the E-double bond of the ring-enlarged lactone
product. Especially valuable in the octenolide family of unsaturated lactones is that this protocol
generates uniquely E-homoallylic lactones, whereas the corresponding Z-homoallylic octenolides
are thermodynamically more stable.10–12 For comparison with Crotti’s protocol, we treated 2-
cyclohexenone with trimethyltinlithium and then with 4-phenylbutene oxide plus LiClO4 and sep-
arately plus Sc(OTf)3; after lead tetraacetate fragmentation, the corresponding 9-membered
olefinic lactone was isolated in only 20–25% yields (versus the 51% yield in entry 1, Table 1).
In the case of Z-2-cycloheptenone, our n+3 ring expansion protocol produced a 10-membered
lactone product (Table 1, entry 7). Reduction of the lactone double bond in this case formed fra-
grant natural product ( )-phoracantholide13 (4) in only three steps and in 45% overall yield from
commercial 2-cycloheptenone (Eq. (3)). This protocol represents the shortest of the many
syntheses of phoracantholide.13 Since both enantiomers of propylene oxide are commercially
available, enantiomerically pure (+)- and (−)-phoracantholide can be prepared via this synthetic
methodology.

9657
Table 1
n+3 Ring expansion reactions via Eq. (2)
O
7
O
1) LiSnMe3
Pb(OAc)4
CaCO3
H2
O
10
(3)
(Ph3P)3RhCl
2)
, BF3-OEt2
4, 45%
O
-78 °C,1h
Variations of our protocol led to the following observations: (1) several epoxides failed to
undergo this n+3 ring expansion protocol—cyclopentene oxide, cycloheptene oxide, cyclooctene
oxide, 1,5-cyclooctadiene mono-epoxide, cis and trans-stilbene oxides, and styrene oxide; (2)
using either tri-n-butyltinlithium or triphenyltinlithium instead of trimethyltinlithium for the
initial Michael addition seriously inhibited subsequent epoxide opening; (3) performing the
epoxide-opening step by adding neat BF3·OEt2 to a mixture of the enolate plus epoxide gave
similar results to those obtained by adding an ether-diluted solution of BF3·OEt2.
The one-flask, multicomponent, domino14 reactions described here can be used also to form
regiospecifically polyfunctionalized acyclic 1,2-disubstituted alkenes. For examples, in only three
steps from Z-2-cyclohexenone, E-homoallylic acyclic diols 5 were generated, with 5a and 5b
being trifunctionalized alkenes (Eq. (4)).

9658
R1
HO
HO
R1
R2
5a,
5b,
5c,
R1 = CH2CH2Ph
R2 = H,
R1 = (CH2)3NO2
R2 = H,
O
O
O
99%
63%
R2
(4)
i-Bu2AlH
R1 = R2 = Me, 99%
In summary, our continuing interest in one-flask, multicomponent, annulation reactions for
efficient and rapid increase in molecular complexity7,15 now includes n+3 ring expansions
featuring intermolecular coupling of ketone enolates with epoxides.16 Further applications of
this methodology, including synthesis of more complex medium-ring lactones and macrolides,
are being pursued.
Acknowledgements
We thank Johns Hopkins University for seed support and the NSF for financial support of
this project.
References
1. Stowell, J. C. Carbanions in Organic Synthesis; Wiley-Interscience: New York, NY, 1979; Chapter 5.
2. Caruthers, W. Some Modern Methods of Organic Synthesis, 3rd ed.; Cambridge University Press: Cambridge,
UK, 1986; Chapter 1.
3. Taylor, S. K. Tetrahedron 2000, 56, 1149–1463.
4. Crotti, P.; Di Bussolo, V.; Favero, L.; Pineschi, M.; Pasero, M. J. Org. Chem. 1996, 61, 9548–9552.
5. Crotti, P.; Di Bussolo, V.; Favero, L.; Macchia, F.; Pineschi, M.; Napolitano, E. Tetrahedron 1999, 55,
5853–5866, and references cited therein.
6. Hesse, M. Ring Enlargement in Organic Chemistry; VCH, Weinheim: Germany, 1991.
7. Posner, G. H.; Webb, K. S.; Asirvatham, E.; Jew, S.-S.; Degl’Innocenti, A. J. Am. Chem. Soc. 1988, 110,
4754–4762.
8. Nakatani, K.; Isoe, S. Tetrahedron Lett. 1984, 25, 5335–5338.
9. Krow, G. R. Org. React. 1993, 43, 251–798.
10. Cope, A. C.; Moore, P. T.; Moore, W. R. J. Am. Chem. Soc. 1959, 81, 3153.
11. Kling, M. R.; McNaughton-Smith, G. A.; Taylor, R. J. K. J. Chem. Soc., Chem. Commun. 1993, 1593–1595.
12. Wang, T.; Chen, J.; Landrey, D. W.; Zhao, K. Synlett 1995, 543–544.
13. Suginome, H.; Yamada, S. Tetrahedron Lett. 1985, 26, 3715–3718, and references cited therein.
14. Tietze, L. F. Chem. Rev. 1996, 96, 115–136.
15. Posner, G. H. Chem. Rev. 1986, 86, 831–844.
16. A typical gram-scale experimental protocol follows: An oven-dried 250 mL round-bottomed flask equipped with
a stir bar was charged with 45 mL of dry, distilled THF. Hexamethylditin (3.02 mL, 4.77 g, 14.6 mmol) was
added via syringe. A yellow color developed upon the addition of methyllithium (9.66 mL of a 1.4 M solution
in diethyl ether). The reaction mixture was then left to stir for 25 minutes at which point it was cooled to −78°C.
Cyclohexenone (1.01 mL, 1.00 g, 10.4 mmol) was added via syringe; cyclohexene oxide (1.26 mL, 1.23 g, 12.5
mmol) was added via syringe followed by the dropwise addition of neat boron trifluoride–diethyl etherate (1.58
mL, 1.77 g, 12.5 mmol). After 5 min at −78°C, the reaction was quenched with 20 mL of distilled water, and
washed twice with 100 mL of distilled water. The organic phase was dried over magnesium sulfate. The resulting
crude material was reconcentrated once from 10 mL of benzene to remove moisture. An oven-dried 100 mL
round-bottomed flask equipped with a stir bar was charged with calcium carbonate (1.56 g, 15.6 mmol) and lead
tetraacetate (5.99 g, 13.5 mmol), equipped with a condenser, and charged with 25 mL of anhydrous benzene. This
mixture was heated to reflux and a yellow color developed. The crude hydroxy ketone was dissolved in 12 mL

9659
of anhydrous benzene and added via cannula at which point the yellow color faded and a large amount of white
precipitate was formed. The reaction was left to stir at reflux for 30 minutes, quenched with 20 mL of distilled
water, and diluted with 50 mL diethyl ether. The ethereal layer was washed with 150 mL of 0.3 M hydrochloric
acid and 50 mL of saturated sodium bicarbonate. The ethereal layer was dried over magnesium sulfate, filtered
and concentrated using rotary evaporation. This mixture was then purified using silica gel column chromatogra-
phy (25% diethyl ether, 75% petroleum ether) to give 970 mg of lactone 3 as a pale yellow oil (50% yield). IR
(neat) 2937 cm−1, 2861, 1715 (s, CꢁO); 1H NMR (400 MHz CDCl3) l 5.53 (quintet, J=4 Hz, 1H), 4.98 (dd,
J=15.6, 9.6 Hz, 1H), 4.42 (m, 1H), 2.29–1.22 (m, 15H); 13C NMR (100 MHz CDCl3) l 178.07, 134.14, 134.09,
78.92, 52.03, 33.35, 31.04, 29.25, 27.15, 26.29, 25.34, 25.18; HRMS [M+] calculated for C12H18O2 194.1307,
observed 194.1313. It is important to note that despite their size and molecular weight, these lactones are volatile;
therefore care should be used when concentrating them.
.